Frontiers in Heat Pipes (FHP), 4, 023003 (2013)

DOI: 10.5098/fhp.v4.2.3003

Global Digital Central

ISSN: 2155-658X

Frontiers in Heat Pipes

Available at www.Ther malFluidsC entral.org

BOILING AND CONDENSATION HEAT TRANSFER COEFFICIENTS

FOR A HEAT PIPE HEAT EXCHANGER
*

R. Laubscher, R.T. Dobson

Department of Mechanical and Mechatronic Engineering, University of Stellenbosch,

Stellenbosch
Private Bag X1, MATIELAND 7602, South Africa

ABSTRACT
This paper reports on the experimental determination of the boiling and condensation heat transfer coefficients for the workin g fluid of a novel
heat pipe heat exchanger at high operating temperatures. This technology is of importance in the high -temperature gas cooled nuclear power and
process heat industries as it eliminates the probability of tritium finding its way into the product stream. The high temperature wor king fluid used
to evaluate the heat pipe heat exchanger was Dowtherm-A, a special high temperature eutectic mixture of 26.5% diphenyl and 73.5% diphenyl
oxide. A specially designed experimental heat pipe heat exchanger was built using a hot gas stream of up to 600 C as the heating stream and col d
water as the cooling stream. The basic features of the design are given. The experimental model is used as a proof of concept and to validate the
theoretical heat transfer modeling analyses. Existing heat transfer coefficients correlations were considered; they include the Rohsenow, Stephan
and Abdelsalam, Gorenflo, and Kutateladze correlations. None of these correlations gave satisfactory results and as such experimentally
determined boil ing and condensation correlations were determined and which gave reasonable results.
Keywords: Heat pipe heat exchanger, Dowtherm-A, heat transfer coefficients, boiling, condensation.
vaporization is transferred with essentially no temperature drop. [This
heat transfer is thus reminiscent of a closed two-phase thermosyphontype heat pipe.] The need for an extra heat exchanger, circulating pump
and pipe work and high pressure working fluid container is thus also
avoided.
Note that in Fig. 1 both the hot stream and the cold stream
are shown as single tubes passing through the container. In an
actual practical case there will be a large number of tubes and these
tubes will not pass straight though the container. It is more likely that
the hot and
cold stream fluids will pass through a two-tube-pass heat
exchange single shell (for the container) as in conventional
tubular heat exchangers technology. The actual design would have to
take careful cognisance of differential thermal expansion stresses,
strains and deformation. The reason the concept has only two single
tubes is for

1. INTRODUCTION
Operating at fuel temperatures in excess of 950 C, as most very high
temperature gas-cooled nuclear reactors are expected to do,
the formation of tritium in the reactor and its occurrence in the
coolant stream is of concern. At these high temperatures tritium
3
(H ) has a particularly high diffusion rate through steel (Jousse, 2007)
and may conceivably find its way into the process stream and
ultimately into a
consumer product. It has been suggested that this concern might
be addressed by including an additional heat transfer loop between
the primary reactor coolant loop and the process heat steam loop. Such
an additional heat transfer loop would require an additional
heat exchanger, additional pipework and circulation pump and a high
temperature and pressure gas or low pressure but high
temperature
liquid metal as the working fluid. Because of the additional
heat
exchanger there will, needs be, also be an additional degradation of the
available process stream temperature. A novel natural circulation twophase thermosyphon-type heat pipe heat exchanger (HPHE) has been
suggested by Dobson and Laubscher (2013). A concept drawing of
such a heat transfer device is given in Fig. 1.
The concept HPHE shown in Fig. 1 consists essentially of a
working fluid container more-or-less half charged with a liquid in
thermal equilibrium with its vapour; and by a judicious choice of
working fluid, at a relatively low pressure, say 0.5 to 1.5 kPa. The
high pressure hot fluid stream is channeled through the evaporator
section, whilst the stream being heated is channeled through highpressurewithstanding tubes positioned in the vapour of the condenser
section.

(Condenser section)
Vapour
Cold
A
stream
inlet

Working fluid container

Cold stream outlet

Hot
stream

Th
e
wo

rkin
g
flui

d boils in the evaporator and condenses as

a film on

Frontiers in Heat Pipes (FHP), 4, 023003 (2013)

DOI: 10.5098/fhp.v4.2.3003

outlet

Liquid A

the colder tubes, collects and forms droplets which, under the
influence
of gravity, fall back into the liquid pool. In this way the latent heat
of
------------------------------------------------------*
Corresponding author email: rtd@sun.ac.za

Evaporator section

Global Digital Central

ISSN: 2155-658X

Hot stream
inlet

Section A-A

Fig. 1 Concept drawing of a novel heat pipe heat

exchanger (HPHE).

ease of experimental data capture and analysis, as single tube heating

(pool boiling) and condensing (film-wise) geometries are well
documented in the published literature.
Although the indicated HPHE working fluid for a 950 C primary
heating stream is sodium, it was deemed expedient from a safety
and
practical point of view to first conduct tests at an intermediate
temperature. The highest non-toxic, non-flammable and commercially
available working fluid for this purpose is Dowtherm-A. It is a eutectic
mixture of 20.5% biphenyl and 70.5% diphenyl oxide, and is
supplied by the Dow Chemical Company (1997). Its properties are
given in Table 1 where it is seen that its boiling point at one
atmosphere is
257.1 C; and it is stored, shipped and suplied in conventional 210
litre
(55
US
gallons)
standard
steel
drums
(or
so-called
barrels).
Unfortunately however the long-term stability of Dowtherm-A has
as yet not been fully established and demonstrated.
A theoretical heat transfer model for a HPHE was developed
(Laubscher, 2013), but the model then needed to be validated. To
do this an experimental apparatus was designed, constructed and tested.
To experimentally validate the theoretical model the boiling and
condensation heat transfer coefficients for Dowtherm-A were needed. It
thus follows that the objectives of this paper may be summarised
as
follows:
Document and evaluate the existing and available boiling
and condensation heat transfer coefficient correlations that
could
be applicable to Dowtherm-A.
Observe the general flow behavior of the working fluid in
the container by means of direct visual observation
through
windows in the sides of the experimental test apparatus.
Also
to check whether the vapour-liquid flow occurs in moreor- less vertical planes along the length of the tubes; as this
was assumed to be the case in the theoretical model.

Evaluate the applicability of the existing correlations for

Dowtherm-A as working fluid, and if needs be generate a new
and better correlation.
To meet these objectives a literature study was undertaken in section
2 in order to identify appropriate boiling and condensation
correlations for inclusion into the theoretical
model. The
experimental work that
was undertaken, in which a Dowtherm-A charged, HPHE was built and
tested is described in section 3. Based on the experimental test
results two new correlations were determined in section 4; one for the
boiling
and the other for the condensation heat transfer coefficients.
An evaluation
of the units
experimental
Property of the accuracy and applicability
value
and
Freezing
point
12 and C
theoretical
correlations
was done and briefly discussed
Boiling
point
at
1
atmosphere
257.1
C
conclusions are drawn in section 5.
Flash point
113
C
Fire point
118
C
Auto-ignition
point
599
C
Table 1 Thermal properties of Dowtherm-A (Dow Chemical
3
Density at1997).
20 C
1056
kg/m
Company,
Critical temperature
497
C
Critical pressure
313.4
kPa
Heat of combustion
30053
kJ/kg
Latent heat at 100 C
345
kJ/kg
Dynamic viscosity at 25 C
3.71
mPas
Thermal conductivity of vapour at 25 C 0.0081 W/mC
Thermal conductivity of liquid at 25 C
0.1379 W/mC
Saturated temperature at 50 kPa
226.67
C

2. LITERATURE STUDY
When evaluating the thermal characteristics of a two-phase closed
thermosyphon-type heat pipe it is important to consider the heat
transfer coefficients between the internal working fluid and the wall
surface in both the evaporator and the condenser sections (Pioro,1998),
where the heat transfer coefficient is defined as
(1)
(

2.1 Boiling Heat Transfer Coefficient

For boiling, the heat transfer coefficient depends on the boiling
regime, of which there are four:
natural convection, nucleate,
transition and film boiling. Of the four, nucleate boiling is the most
desirable boiling regime in practice because high heat transfer rates
can be achieved in this regime and also with a relatively small
temperature difference between the heated surface and liquid (Cengel
2006). For nucleate boiling, the heat flux is the dominant factor
influencing the heat transfer coefficient. The heat transfer rate is also
influenced by the physical properties of the working fluid, which may
even vary along the axial length of the heat exchanger. In general,
nucleate boiling heat transfer coefficient correlations contain a wall
heat flux, a temperature term to capture the fluid pressure and density
and a fluid property dependent factor, such as the fluid Prandtl number,
to capture the effect of thermal conductivity, surface tension and heat
capacity (Pioro et al. 2004). Thome (2006) stated that experimental
results for heat flux
and wall superheat T typically fit to an
exponential
equation
of one
of the following forms,
or
, where n equals 3, 2 or 0.7, respectively.
The Rohsenows (1962) boiling heat transfer correlation is
the most widely used and can be used for any geometry, since it was
found that the rate of heat transfer during nucleate boiling is essentially
independent of the geometry and orientation of the heater surface;
rather, it is more dependent on the bubble pumping. The
Rohsenow
correlation is

(2)
The results obtained using the Rohsenow correlation can be out by
about 100% for the heat transfer rate, given a temperature
difference between the saturated liquid and the heated surface (or
the excess
temperature); and given the heat transfer rate may be out by about 30%
for the excess temperature. For Dowtherm A, the constant Csf is
0.00695 and n is 1.72 (Jouhara, 2005).
Stephan and Abdelsalam (1980) formulated a correlation
containing four groups of variables using regression techniques
for water, refrigerants, organics and cryogens as
[(

)]
[

]
(3)

where
by

is the bubble departure diameter, and is given

(4)
where
is the contact angle that may be assigned a fixed value of
35,
irrespective of the fluid, and aL is an an experimentally determined
parameter. This correlation is for organic fluids and it could be
expected that it would also generate accurate results for Dowtherm-A.
Gorenflo (1993) proposed a fluid-specific correlation and
included the effect of surface roughness and reduced pressures and
contains a

working fluid specific and a reference heat transfer coefficient

and is given by
(

(5)
where the pressure correction factor
by

is given
(6)

and
(7)
where
is the surface roughness and is set at 0.4 m when unknown.
This correlation is applicable for a reduced pressure range from
about
0.0005 to 0.95, and is known to be applicable over a wide range of
heat
fluxes and pressures and according to Thome (2006) is also very
reliable.
The Kutateladze et al. (1961, 1966) correlation is applicable to a
wide range of conditions (fluid type, heater geometry and surface
roughness) and is given by
(

heats the liquid and boiling occurs. The so-formed vapour fills
the vapour section above the liquid. The vapour then condenses on
the condenser tubes and heat is in turn then transferred from the
condenser wall, through the tube wall and into the cold stream liquid.
The HPHE consists of a container for the working fluid, a 101.6
mm diameter evaporator tube and a 21.3 mm diameter condenser
tube, all of 316L stainless steel, and two large glass windows. Also
shown in Fig. 2 is the pipework, flow control valves, pressure
relief valve, constant head water supply tank, working fluid sump
and vacuum pump. (Not show is the continuous combustion chamber
hot gas supply and the mild steel support structures.) K-type
thermocouples were used for all the temperature measurements. The
thermocouple junctions were positioned at four vertical-plane locations
along the axial length of the pipes, 0.058, 0.248, 0.428 and 0.618 m
from the hot stream inlet side. For each vertical plane three
temperatures measurements were taken around the periphery of the
heat exchange pipes, in the liquid next to and just above the
evaporator pipe, and two above each other in the vapour space; giving
a total of 44. Four thermocouples were used to measure the inlet and
outlet temperatures of the heating gas and cooling water. The room
temperature was also measured. Mass flow rate of the combustion
gases was determined using a bell-mouth located at the gas- burner
inlet to which was added the liquefied petroleum gas (LPG) fuel, as
measured using an orifice-plate type flow meter. The mass flow rate
of the cold water stream from the constant heat tank was determined
using a measuring-cylinder and stop-watch.

(8)

2.2 Condensation Heat Transfer Coefficient

The laminar film condensation heat transfer coefficient is determined
theoretically by using a force and energy balance over a liquid element
within the condensate film flowing around the condenser tube,
rather than by generalisations of experimental data (Mills, 1995).
It is dependent on the thermal conductivity of the working fluid
and the thickness of the condensate film on the condenser tube and is
given by
(9)
The film thickness is a function of the Reynolds number, liquid and
vapour densities, gravity and angle of location on the condenser wall
and is given by
(

(10)

where the Reynolds number is given as

[

Working fluid container

Integrating around the periphery of the condenser tube, the average

heat transfer coefficient for laminar film condensation of a horizontal
tube can be derived, as

(11)

(12)

3. EXPERIMENTAL WORK
An experimental heat pipe heat exchanger was designed, built and
installed in the laboratory, as shown in Figure 2. As well as allowing
for the experimental determination of the heat transfer rates and the
boiling and condensation heat transfer coefficients, it also has windows
in the front and backside thus allowing visual observation of internal
working fluid behaviour. A hot stream flows through the evaporator
pipe and

Fig. 2 Schematic layout of experimental setup. The dimensions

of the working fluid container are (excluding the flanged
instrumentation and observation platforms) 596 mm
high,
671 mm wide and 516 mm deep. The hot gas tube diameter
is 101.6 mm, whilst that of the cold water pipe in the
vapour
space is 21.8 mm.

4. RESULTS AND ANALYSIS

During experimentation visual inspection was done on the boiling and
condensation processes by observing the evaporator and condenser
through a glass window, which was incorporated into the experimental
design. Figure 3 shows a photograph of the HPHE evaporator section
whilst boiling. When observing the evaporator during operation
the following is noted: the working fluid flows mainly radially away
from the evaporator surface and the boiling frequency (bubble
formation and

separation) is observed to be almost constant. Figure 4 shows a

photograph of the condenser section during operation. When observing
the condensation process the following is noted: all along the length of
the condenser droplets forms at the bottom of the condenser tube and
when the droplet is large enough it detaches itself from the tube
surface
and falls into the working fluid pool. It is also noted that there is
no
droplet formation on the remaining periphery of the condenser
tube. The more detailed boiling and condensation experimental
procedure will be processed and discussed next.

4.1 Boiling Heat Transfer Coefficient

The experimental boiling heat transfer coefficient was determined by
dividing the experimentally determined evaporator wall heat flux
with the difference between the experimentally measured difference
between the saturated working fluid temperature and the average
wall
surface temperature (
(

(13)

) as

where
(19)
is the average vapour temperature in the heat exchanger and
is the average condenser tube surface temperature. In figure 7 the
condensation heat transfer coefficient as calculated by equation 18
is seen to be more or less constant for all the experimental value from 0
to
5000 W/m2K. Assuming a correlation of the
form
the constants C, m and n were also determined using
multivariable linear regression to obtain a correlation with
of the form
(20)
Figure 8 shows that equation 18 better predicts the condensation
heat
transfer rate than does the so-called Nusselt theory as given by
equation
12, but with a relatively wide scatter band of
70%.

where
(14)
and
is the
the
heat exchanger.
where
is theand
average
ofaverage
all the evaporator
thermocouple
temperature
readings
liquid poolsurface
temperature
in (15)
Figure 5 shows the experimentally determined boiling heat
transfer
Working fluid flow
Hot gas flow
direction
direction
Fig. 3 Photograph of evaporator tube.

coefficients as a function of the wall heat flux; also shown are the four
correlations mentioned in the literature study. It is seen that the
average experimentally determined heat transfer coefficient is
about 5000
2
1000 W/m K and that it is relatively insensitive to the heat flux. All
the correlations gleaned from the literature are all seen to under-predict
2
the experimental values by roughly 3000 W/m K or about 70%. As
this is a relatively large discrepancy an experimentally determined
heat transfer
coefficient
correlation
of
form
was
considered.
The liquid-surface temperature difference
is
captured by
, the ability of the heat to diffusion through the
liquid by the
, and the fluid properties by
. The constants C,
m, n and z were determined using multivariable linear regression to
give an experimentally determined boiling heat transfer coefficients of
(16)

The experimental condensation heat transfer coefficient was

determined using

Fig. 6 shows that equation 16 predicts the boiling heat

transfer
coefficients to within 10% of the experimentally determined
boiling
heat transfer coefficients.

(17)
where

4.2 Condensation Heat Transfer Coefficient

(18)

Cold stream inlet

Condensate droplets forming

Condenser tube
Fig. 4 Photograph of condensation process.

5. CONCLUSIONS
The boiling heat transfer coefficient of the heat pipe heat exchanger
using Dowtherm A was experimentally determined and it was found
that the correlations ascribed to Rohsenow, Gorenflo, StephanAdelsalam, and Kutateladze and as given by equations 2, 3, 4 and 5,
respectively were all some 70% too low. The experimentally
determined correlation for boiling heat transfer coefficient as
given by equation 16, however, captured the experimental values to
within
10%.

increases. Also liquid tends to collect at the ends of the

thermocouple probes measuring the vapour temperature.
A large amount of carry-over of liquid out of the liquid pool (as
the vapour escapes out of the liquid pool) was not noticed.
However, as
intimated in the objectives in the introduction, the does indeed
tend
to flow as vapour upwards and liquid downwards in more-or-less
vertical planes along the length of the tubes with apparently a
minimum of flow parallel to the direction of the heating and
cooling tubes. The assumption uniform upwards and downwards
flow in formulating the theoretical model (Dobson and
Laubscher, 2013; Laubscher, 2013) was thereby indeed validated.

NOMENCLATURE
A
a
cp
Csf
D
D

Area, m
Dimensional experimental variable
Specific heat, J/kg K
Rohsenow constant
Diamter, m
Diamter, m

hb (W/m K)

Boiling heat transfer coefficient,

The theoretically determined condensation heat transfer coefficient

assuming film-wise condensation theory and as was calculated
using equation 12 did not really correlate with the experimentally
determined heat transfer coefficient values. As shown in Fig. 7, for
experimentally determined values ranging from 1000 to 5000
2
W/m K, the theoretical value remained more-or-less constant
at
2
1400 W/m K. From this it is concluded that traditional film-wise
condensation theory should not be automatically be assumed to
apply in the case of the condensation process as exhibited in the
heat
pipe heat exchanger.
The experimentally determined condensation heat transfer
coefficient correlation as given by equation 20 did indeed
capture
the average experimental values as seen in figure 8 but then only
to
within 70% of the experimental
values.
The large scatter of the experimental condensation heat transfer
coefficient results as exhibited in Fig. 8 can be ascribed to
the
following affect: as the droplets form at the bottom of the
condenser
tube the film thickness increases and therefore the condensation
heat transfer coefficient value drops. When the droplet detaches
itself from the condenser surface the film thickness suddenly
decreases
and the condensation heat transfer coefficient value therefore
rapidly

Evaporator wall heat flux, q (W/m )

in

Predicted boiling heat transfer coefficient,

2
h b (W/m K) using equation 10

Fig. 5 Experimental and sundry boiling heat transfer coefficient

corr el ations as a function of the evaporator wall heat
flux.

Fig. 7 Theoretical condensation heat transfer coefficient

equation using Nusselt theory as a function of the
experimental heat transfer coefficient results.

Experimental boiling heat transfer coefficient,

2
hb (W/m K)
Fig. 6 Predicted boiling heat transfer coefficient (using
equation 16) as a function the experimentally
determined heat transfer coefficient.

Fig. 8 Correlated (equation 19) condensation heat

transfer coefficient equation as a function of the
actual experimental heat transfer coefficient results.

FPF
g
h
hfg
k
L
nf
N
P
Pr
Pcrit
Pr

Re
Rp
r
T

Factor
2
Gravity, m/s
2
Heat transfer coefficient, W/m K
Latent heat of vapourisation, J/kg
Thermal conductivity, W/m K
Length, m
Characteristic length, m
Mass flow rate, kg/s
Gorenflo coefficient
number (of tubes)
Pressure, Pa
Reduced pressure
Critical pressure, Pa
Prandtl number,
Heat transfer rate, W
Heat flux, W/m2
Reynolds number, Re = VD/
Surface roughness, m
Radius, m
Temperature, C

Greek symbols
2

Heat transfer coefficient, W/m K

(Film) thickness, m

angle, or rad.

Surface tension, N/m

Density, kg/m3

Viscosity kg/m.s
2

Kinematic viscosity = / , m /s
Subscripts
avg
b
C
c
cold
cond
evap
hot
in
inlet
int
l
o
out
outlet
s
sat
v

average
boiling
constant
condensing
Cold stream
Condenser
Evaporator
Hot stream
Into section
Inlet to HPHE
Internal
Liquid
Outside
Out of section
Out of HPHE
Surface
Saturation
Vapour

REFERENCES
Cengel, Y., 2006, Heat and Mass Transfer, McGraw Hill, New York.
Dobson, R.T. and Laubscher, R., 2013, Heat Pipe Heat Exchanger
for
High Temperature Nuclear Reactor Technology, Proceedings of
the

11th International Heat Pipe Symposium (11th IHPS), June 9-12,

Beijing, China.
Dow Chemical Company, 1997, Dowtherm A Heat Transfer Fluid:
Product technical data. [Online] Available at:
http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_0030/0
9
01b803800303cd.pdf?
filepath=/heattrans/pdfs/noreg/176 01337.pdf&fromPage=GetDoc
Gorenflo, D., 1993, Pool Boiling, VDI-Heat Atlas, Dusseldorf,
Germany.
Jouhara, H., 2009, An Experimental Study of Wickless Miniature
Heat Pipes Operating in the Temperature Range 200 C to 450 C,
Heat
Transfer
Engineering,
30(13),
1041-1048.
http://dx.doi.org/10.1080/01457630902921113
Jousse,
A.,
2007,
Tritium Transport in Very High
Temperature Reactors for Hydrogen Production, Report Number
UCBTH-07-005, University of California, 31 August.
Kutateladze, S.S and Borsnansky, V.M., 1966, A Concise
Encyclopaedia of Heat Transfer, Pergamon Press, New
York.
Kutateladze, S.S, 1961, Boiling Heat Transfer, International
Journal of
Heat
and
Mass
Transfer,
July,
31-45.
http://dx.doi.org/10.1016/0017-9310(61)90059-X
Laubscher, R. and Dobson, RT., 2013, Theoretical and Experimental
Modeling of a Novel Heat Pipe Heat Exchanger for High Temperature
Nuclear Reactors Technology, Applied Thermal Engineering, In print.
Mills, A.F., Heat Transfer, Prentice Hall, New Jersey,
1995.
Mostinskii, I.L., 1963, Application of the Rule of Corresponding
States for Calculation of Heat Transfer and Critical Heat Flux,
Teploenergetika, 4, 66.
Pioro, I.L., 1998, Experimental Evaluation of Constants for the
Rohsenow Pool Boiling Correlation, International Journal of Heat
and
Mass
Transfer,
29,
January,
2003-2013.
http://dx.doi.org/10.1016/S0017-9310(98)00294-4
Pioro, I.L., Rohsenow, W. & Doerffer, SS., Nucleate pool-boiling heat
transfer II: assessment of prediction methods, International Journal
of Heat and Mass Transfer, 23, August 2004, pp. 5045-5057, 2004.
Rohsenow, W., A Method of Correlating Heat Trasnfer Data
for
Surface Boiling of Liquids, Trans. ASME 74, pp. 979-976,
1952.
Stephan, K. and Abdelsalam, M., 1980 Heat Transfer Correlations for
Natural Convection Boiling, International Journal of Heat and Mass
Transfer, 23(1), 73-87.
http://dx.doi.org/10.1016/0017-9310(80)901404
Thome, J., 2006, Boiling Heat Transfer on External Surfaces,
Engineering Data Book III, Wolverine Tube Inc., pp. 9-1 and 9-11.